Predictors of molecular subtypes in women with breast cancer in Rwanda

Abstract

INTRODUCTION: Breast cancer (BC) constitutes a major public health problem worldwide. It remains a major scientific, clinical and societal challenge, generally in Africa and particularly in Rwanda. The purpose of this study was to determine clinical and histopathological predictors of BC molecular subtypes in Rwandan women.METHODS: A retrospective cohort study including patients with histological confirmation of BC. Using R statistical software, a regression model for multinomial responses was developed. Univariate and multivariate logistic regression analyses were used to identify independent BC molecular subtypes predictors. A two-sided p<0.05 indicated a statistically significant difference.RESULTS: Forty seven percent of cases presented with advanced stages (Stage III and IV). Postmenopausal BC (p=0.0142), absence of infertility (p=0.018) predicted Luminal A subtype with a predictive accuracy of 0.65. Age (p=0.003), postmenopausal BC (p=0.005), absence of axillar lymph nodes (p= 0.008) and poorly differentiated tumor (p=0.012) were predictors for Luminal B subtype with a predictive accuracy of 0.86. Age (p=0.045), BMI (p=0.005), rapid progression (p=0.032), tumor size T2-T3 (p<0.001) were predictors of HER2-Enriched subtype with a predictive accuracy of 0.70. Age below 40 (p=0.005), painless mass (p=0.030), nodal involvement (p=0.008), Nottingham grade 3 (p<0.001) predicted Triple Negative tumors with a predictive accuracy of 0.71.CONCLUSION: Clinical and histopathological tumor characteristics can be used to predict BC molecular subtypes with acceptable accuracy. Further studies are needed to explore the possibility of developing a scoring system for clinical decision-making, especially in settings where immunohistochemistry testing is limited.